Scale-Up Methodology for Municipal Water Nanofiltration Based on Permeate Water Quality

Abstract

A semiempirical differential model based on homogenous solution and diffusion is developed to analyze contaminant removal during municipal water nanofiltration at high feed water recoveries. An intrinsic transport parameter derived using this model from bench-top experiments employing flat membrane sheets is used to predict contaminant rejection in large-scale experiments using single and multiple spiral-wound modules conducted under the Information Collection Rule. Because solute back-transport dominated permeate flux, concentration polarization was negligible during these nanofiltration experiments. Permeation coefficients were found to increase monotonically with increasing solute aqueous diffusivity. Pseudostochastic simulations were performed in a purely predictive manner using the Monte Carlo methodology to theoretically calculate permeate concentrations in large-scale facilities using probability density functions for permeation coefficients (obtained from bench-top experiments) and observed feed water concentrations. These numerical simulations revealed that removal of natural organic matter and precursors to trihalomethanes, haloacetic acids, and total organic halide could be successfully scaled up from bench-top tests with flat membrane sheets to large-scale facilities employing spiral-wound elements. Thus, factors that strongly influence natural organic matter removal (steric and polar interactions) appear to be scaled accurately. However, statistically significant discrepancies in inorganic contaminant concentrations between theoretical predictions and observations from pilot- and full-scale tests suggest that the proposed methodology does not accurately quantify electrostatic interactions, ion pairing, and complexation.